Elevating predictive reliability: time-varying parameter bayesian deep learning techniques for flood probability forecasting

doi:10.1016/j.jhydrol.2025.134597

Elevating predictive reliability: time-varying parameter bayesian deep learning techniques for flood probability forecasting

2025-11-10

0

PRE

AI

DOI：10.1016/j.jhydrol.2025.134597

Original

Original request for help

Save

Abstract

En 中文

• A novel F-TV-BLSTM model is proposed to enable probabilistic flood forecasting. • Fourier basis function is used to characterize time-varying parameter mechanism. • Time-varying parameter Bayesian captures non-stationary rainfall-runoff relationship. • Integrating precipitation forecasts from FourCastNet boosts flood forecasting accuracy.

Elevating predictive reliability: time-varying parameter bayesian deep learning techniques for flood probability forecasting

Abstract

Journal

Organization